Emergency contraceptive pills (ECPs) can substantially reduce the chance of unintended pregnancy after unprotected intercourse. However, over the longer term, broad availability of this method could affect women's use of other contraceptive methods, which could in turn have either beneficial or adverse consequences on the risk of pregnancy and sexually transmitted infections (STIs). Data on this issue are needed to inform policies and practices regarding provision of ECPs to women. The primary purpose of the proposed randomized trial is to evaluate the impact of maximally increased access to ECPs on pregnancy and STI rates. Secondary research aims are (1) to determine what contraceptive or STI-prevention behaviors ore influenced by ready access to ECPs. and (2) to examine determinants of these behaviors. The trial will be conducted in young, primarily minority women, a population with a disproportionately high risk of these two outcomes. Sexually active women who are using barrier methods of contraception, oral contraceptive pills, or no contraception will be enrolled at two clinics in Indiana and California. Each woman will be randomly assigned to one of two groups. In the standard care group, women will be counseled about ECPs and invited to come to the clinic to obtain them (at usual clinic charges) when necessary. In the advance provision group, women will be given a supply of ECPs free of charge to keep at home in case of need. The two groups will be monitored over the following year. The primary analysis will compare the two groups for: 1. the 12-month incidence of pregnancy, and 2. the combined 12-month incidence of three STIs: cervical gonorrhea infection, cervical chlamydia infection and vaginal trichomoniasis. In addition, we will also compare behaviors in the two groups, specifically use of condoms, other contraceptive methods, and ECPs, and also motivating or deterrent determinants of these behaviors using a modified Health Behavior framework. These secondary analyses should help explain the primary biologic outcomes and should be useful in the development of counseling messages and service protocols in the future.